1. Field of the Invention
The present invention relates to a recovery method for a ring-based network, in particular to the recovery method for a ring-based network that uses a link as the connection base.
2. Description of the Related Art
The form of connecting networks includes Line, Ring, Bus, Star, Mesh or Tree topologies, wherein each node in a ring-based network is serially connected to form a closed ring structure. In local area network or industrial network applications, each node has a network switch, and each switch is connected to a communication device such as a computer, a server or a terminal for providing an information exchange function. In addition, the circuit of the ring-based network is a closed circular loop, so that when old information is looped and transmitted continuously and infinitely in the ring-based network, the bandwidth of the network will be occupied by old information, and new information cannot be transmitted. As a result, the network is paralyzed to form a broadcast storm. To solve this problem, the ring-based network generally adopts a ring port of a certain switch set at a blocked status, so that only specific control packets can be passed, and data packets are prohibited to avoid the problem of having information to be looped in ring-based network infinitely.
In addition, if any node or link in the network is abnormal, damaged, or interrupted, and the network cannot be used, the network will be paralyzed due to the information transmission characteristic of the ring-based network. To assure a smooth communication in the ring-based network, related solutions have been provided by network equipment suppliers and manufacturers, such as using an algorithm of the rapid spanning tree protocol (RSTP) to increase the recovery rate, but the reduced recovery time of the aforementioned solution still cannot meet the industrial requirements.
Among various solutions, the media redundancy protocol (MRP) set by the IEC 62439 standard of the International Electronics Commission (IEC) is the most eye-catching one. The MRP sets a ring portion of a master switch to a blocked status, and periodically sends out a detection signal to detect the working status of each switch or each link. If a certain switch or link in the loop is damaged, the ring port of the master switch will be switched from the blocked status to a forward status immediately to allow specific control packets and data packets to pass through simultaneously and start and conduct a spare link or network, so that the data packets can be transmitted through the spare link or network to prevent the termination of data transmissions or data loss caused by the interrupt of the network. After the damaged switch or link is repaired, the ring port at the forward status is switched back to the blocked status to enhance the stability of the ring-based network. However, the aforementioned method may cause misjudgment and failure when the network is overloaded or the range of the network topology is too large. Therefore, the IEC 62439 specification recommends that the ring-based network preferably uses less than 50 nodes, but such requirement cannot meet the actual requirement of the present industrial networks.
In conventional ring-based network recovery methods, a detection signal is sent out periodically and a timeout trigger mechanism is adopted for disconnection detections, but a misjudgment may occur easily under the aforementioned conditions. The recovery time of the ring-based network is set when the recovered control packet is informed, and the recovered control packet is looped around the whole ring-based network. After each node received the recovered control packet, each node will clear its own filtering database (FDB) to fit the ring topology after the link down occurs. However, the aforementioned method can recover the packets only after the last node has completed clearing its own filtering database, so that the recovery efficiency of the ring-based network is low.
In addition, if the ring-based network based on the master-slave switch architecture has a plurality of master switches, it is necessary to use a complicated and time-consuming priority setting procedure to determine using the switch of a second priority to form a master switch, when the master switch is damaged, so as to avoid a direct paralysis of the network. If the network has only one master switch, a serious single point failure problem exists.